Lifen
Lifen sends medical documents and files them in patient records for French hospitals, clinics and independent clinicians, with Lifen Documents and Lifen Intégration running in the same application. Lifen Documents reads a report to detect its recipients and patient, then sends it through MSSanté secure messaging, Apicrypt, a secure patient link or postal mail, drawing on a directory of French health professionals, and can copy it to the patient's national shared record. Lifen Intégration takes incoming reports, scans, secure email attachments and device output such as ECGs, uses its AI to extract the patient's identity, the visit, the date and the document type, and files the document in the hospital information system. Documents with complete information can be filed fully automatically under each department's settings, and the rest go to a work queue for a two click check. The company says 94% of documents match to the patient identity automatically and 85% are integrated without human intervention, and a Visa Médecin module adds physician sign off for secretariats.
More than 4 million documents a month pass through the platform. CHRU Nancy's central document team files 50,000 to 60,000 items a month into DxCare, CH Soissons uses it with Easily, and facility customers include Gustave Roussy, Institut Curie, CHU Strasbourg, Ramsay Santé and Almaviva Santé.
Founded in 2015 in Paris, Lifen counts more than 800 hospitals and 20,000 independent professionals as users. Receiving documents is free, while sending and record integration are sold by agreement and through a French public purchasing body. Data is hosted in France under the HDS health data hosting framework.
Capability Axes
An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check. How grades read
Lifen's document models identify sender, recipient, patient, encounter, date and document type from reports and incoming paper, and those fields drive both directions of the product: routing outbound correspondence to the right channel and filing inbound items in the right place in a patient record. Around the models sit secure delivery, a national directory of professionals, receipt and work queues, which have value for practices exchanging documents even when a person picks the type.
The AI is what turns unstructured clinical and administrative documents into the structured identity and encounter data that let 85% of documents file without a person, so it is the engine of the integration module inside a broader exchange service.
The integration queue separates documents with complete extracted information from those that need checking. When a match needs correcting, staff update the fields and click to file, and support guidance divides document types into ones the classifier handles reliably, ones it checks more often and ones left for manual selection.
Departments can switch on fully automatic filing for ready documents, and the company reports that 85% of integrated documents reach the record without human intervention. The integration page promises to eliminate identity verification and document storage errors in the EHR. That absolute claim covers the automatic path, where a document is filed in the chart with no person seeing that particular match first.
The document model is proprietary and built in France. It identifies patient and encounter information, detects document type and uses each recipient's channel preference for outgoing reports. Support documentation sorts document types into groups the classifier recognizes reliably, handles with more frequent checks or leaves for manual selection, and new types join the last group until the model has enough examples. Release notes cover changes to the supported workflow. The deployed architecture, a version tied to each training update and a breakdown of the training corpus are not given, so a hospital cannot tell when a model change might alter matching behavior.
The document models run in house, with no dependence on an outside AI service, which keeps the model layer inside the company. A subprocessor list linked from the privacy materials sets out providers by activity, data type and location, naming cloud hosting on AWS Paris, S3NS and Azure among others, and outside services such as the printer used for postal delivery.
Delivery also runs through public channels, MSSanté and the national shared record, which operate under French state rules rather than as Lifen subprocessors. That maps most of the processing chain. Transfer and retention terms at each provider are not stated, and nothing describes how operational logs from the document workflow move between supporting providers.
CHRU Nancy's central document team files 50,000 to 60,000 items a month into DxCare and kept up throughput as staffing fell, according to the company's 2025 account, and CH Soissons uses the same integration with Easily for paper, secure email attachments and ECGs. Across the platform, more than 4 million documents pass through each month, and more than 5 million were integrated into hospital records in 2024.
The company also reports 94% automatic identity matching and 85% integration without human intervention, without a sampling method or a measure of wrong patient filings among the automatic ones. DataLab's research results belong to a separate product and say nothing about filing accuracy.
The care terms limit health data use to coordination between healthcare professionals, and the service privacy charter covers downstream access and return or destruction of data after the applicable period. The AI documentation describes pseudonymization and rules out commercial reuse of data. Feedback based learning belongs to the separate research product.
For the care software, nothing says whether corrections made in the filing queue train the document model, and no retention period is given for each model related field. The error and duplicate queue catches rejected items, but there is no playbook for a wider model safety incident, such as a run of documents filed to the wrong patient.
Lifen's privacy charter covers processing health data for medical document exchange and integration under the EU General Data Protection Regulation, identifies Lifen as a processor when it acts for an institution or professional, and sets out customers' instructions and rights in that relationship. Other processors must be bound by contract, with the list of services and locations supplied separately. The platform also operates under France's HDS health data hosting framework. There is no US business associate agreement and no HIPAA positioning, which fits a service sold to French hospitals and clinicians under French and European rules.
The company holds HDS health data hosting certification and ISO 27001 and describes an earlier Bureau Veritas renewal audit for its hosting service, along with encryption in transit and at rest, audit logging and account authentication controls. The service privacy charter lets the data controller request evidence of certification.
No current certificate is available to download for the document modules, and the scope of a current security report for every partner in downstream delivery is not given. The HDS certification covers the hosting operations defined in France's health data framework, which every provider holding French patient data must meet.
The products deliver clinical documents and file them in a record, reading identity, visit and document information to do so. The service includes access to France's MSSanté network and can send reports to the national shared medical record when the hospital has turned that on. These functions are administrative and work on reports a provider has already written. The terms cite French rules on hosting and securely sharing health information. No medical device authorization is claimed for document routing, and the software is not presented as establishing a diagnosis or treatment plan, which matches how it is used.
Predictions and traces in the document workflow are auditable. Support guidance sets out how new document classes are introduced, with manual selection for types outside automated coverage and verification for classes where confidence is lower, and administrators choose which categories are enabled in each department. The company also describes traceability and documentation prepared for the European AI Act. That is a governance structure with named controls.
There is no independent audit of classification behavior and no comparison of wrong patient matches across specialties, document sources or patient groups, such as patients with common names or incomplete identity data.
Staff can correct extracted information in the verification queue and review an error or duplicate status when a downstream system rejects an item, and the privacy charter describes data rights while separating requests that the organization responsible for the record must decide. The service terms limit liability for several kinds of access problems and consequences of using the platform.
No remedy covers a document filed to the wrong patient, no guarantee commits to correct a model classification within a set time, and no right to compensation attaches to a wrong automated match, despite the promise to eliminate filing errors.
Lifen connects over HL7 to the hospital's patient administration system to match identities and sends documents into the electronic record using HL7 ORU or MDM messages. CHRU Nancy files into DxCare and CH Soissons into Easily, and the company supports France's PFI interoperability program for documents arriving by MSSanté and reports from other digital health applications. It also shares documents to the national medical record after looking up the official patient identifier. That establishes named production writes and defined input feeds. Full bidirectional exchange with either named record, reading and updating the same patient fields, is not shown.
The service charter names AWS Paris for storage and states that data can transit through Dublin without being stored there, and it lists S3NS and Azure among other providers, with the subprocessor annex giving each one's role and geographic scope. The care products are hosted software with data held in France under the HDS framework, encryption and strong authentication. Exact retention for each data type is not specified. Customers have no option to run the software on separate infrastructure, and tenant isolation at each step, from classification to delivery and exchange with the receiving record, is not described.
A professional can receive documents through Lifen messaging without paying for an account. Sending is paid in some arrangements, the support center handles account requests by facility or practice, and institutions request a demo for record integration. Lifen is listed with a French public purchasing organization for eligible facilities. No standard fee is given for sending or record integration, whether by hospital size, user or document volume.
The free receipt offer covers inbound messaging only, not the outbound workflow or the work of connecting a record system, so a buyer knows what is free and how the paid parts are bought, but not what they cost.
The document exchange serves independent clinicians, clinics and hospitals sending records to patients and care partners, and the integration product serves hospital document teams such as CHRU Nancy's central office and CH Soissons. Facility customers range from cancer centers, Gustave Roussy and Institut Curie, to university hospitals such as CHU Strasbourg and private groups Ramsay Santé and Almaviva Santé, and the Private Hospital of Côtes d'Armor gives a testimonial.
Document classes include consultation reports, discharge letters and stay reports across departments. Classification and patient matching performance is not compared across document classes or settings, and DataLab's research validation belongs to a separate product.
Alternatives to Lifen
Other products that do the same kind of work as Lifen, and what each one does.
Blooming Health runs outreach, social care navigation and benefits renewal for health plans, health systems, government agencies, area agencies on aging and community organizations.
Dexit is 314e Corporation's HIPAA cloud fax and intelligent document processing product for hospitals, RCM companies and physician practices.
Tennr automates inbound document and referral operations for the organizations that receive them.
Carefeed runs admissions, workforce communication, family engagement and billing for senior living, skilled nursing and long term care organizations.
CertifyOS sells a provider data platform to health plans, health systems and digital health companies.
Aidin runs the discharge referral, from the point a case manager decides a patient needs post acute care to the point a facility accepts them.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
| Entry Price | Pricing Basis | BAA Tier | Implementation | Source |
|---|---|---|---|---|
|
Free to receive documents
|
Free receipt. Professional sending and institutional integration by agreement | French GDPR processor terms. No US BAA tier stated | No integration fee is listed for Lifen Documents or Lifen Intégration. | Vendor Published |
Receiving documents through Lifen messaging is free. Sending may require payment, record integration starts with a demo, and Lifen is listed with a French public purchasing organization. No list prices are given for either paid workflow.